crawl4ai-cloud-sdk 0.3.0

Lightweight cloud SDK for Crawl4AI - mirrors the OSS API

Crawl4AI Cloud SDK for Python

Lightweight Python SDK for Crawl4AI Cloud. Mirrors the OSS API exactly.

Note: This SDK is for Crawl4AI Cloud (api.crawl4ai.com), the managed cloud service. For the self-hosted open-source version, see github.com/unclecode/crawl4ai.

Claude Code Plugin

Use Crawl4AI directly inside Claude Code with 9 built-in tools — no Python needed. See the plugin README for details.

/plugin marketplace add unclecode/crawl4ai-cloud-sdk
/plugin install crawl4ai@crawl4ai-claude-plugins

Installation

pip install crawl4ai-cloud-sdk

Get Your API Key

1. Go to api.crawl4ai.com
2. Sign up and get your API key

Quick Start

import asyncio
from crawl4ai_cloud import AsyncWebCrawler

async def main():
    async with AsyncWebCrawler(api_key="sk_live_...") as crawler:
        result = await crawler.run("https://example.com")
        print(result.markdown.raw_markdown)

asyncio.run(main())

Features

Single URL Crawl

result = await crawler.run("https://example.com")
print(result.success)
print(result.markdown.raw_markdown)
print(result.html)

Batch Crawl

urls = ["https://example.com", "https://httpbin.org/html"]

# Wait for results
results = await crawler.run_many(urls, wait=True)
for r in results:
    print(f"{r.url}: {r.success}")

# Fire and forget (returns job)
job = await crawler.run_many(urls, wait=False)
print(f"Job ID: {job.id}")

Configuration

from crawl4ai_cloud import CrawlerRunConfig, BrowserConfig

config = CrawlerRunConfig(
    word_count_threshold=10,
    exclude_external_links=True,
    screenshot=True,
)

browser_config = BrowserConfig(
    viewport_width=1920,
    viewport_height=1080,
)

result = await crawler.run(
    "https://example.com",
    config=config,
    browser_config=browser_config,
)

Proxy Support

# Shorthand
result = await crawler.run(url, proxy="datacenter")
result = await crawler.run(url, proxy="residential")

# Full config
result = await crawler.run(url, proxy={
    "mode": "residential",
    "country": "US"
})

Deep Crawl

result = await crawler.deep_crawl(
    "https://docs.example.com",
    strategy="bfs",
    max_depth=2,
    max_urls=50,
    wait=True,
)

Job Management

# List jobs
jobs = await crawler.list_jobs(status="completed", limit=10)

# Get job status
job = await crawler.get_job(job_id)

# Wait for job
job = await crawler.wait_job(job_id, poll_interval=2.0)

# Cancel job
await crawler.cancel_job(job_id)

Migration from OSS

Zero learning curve — your existing code works:

# Before (OSS)
from crawl4ai import AsyncWebCrawler
async with AsyncWebCrawler() as crawler:
    result = await crawler.arun(url)

# After (Cloud)
from crawl4ai_cloud import AsyncWebCrawler
async with AsyncWebCrawler(api_key="sk_...") as crawler:
    result = await crawler.run(url)  # arun() also works!

Environment Variables

export CRAWL4AI_API_KEY=sk_live_...

# API key auto-loaded from environment
crawler = AsyncWebCrawler()

Error Handling

from crawl4ai_cloud import (
    CloudError,
    AuthenticationError,
    RateLimitError,
    QuotaExceededError,
    NotFoundError,
)

try:
    result = await crawler.run(url)
except AuthenticationError:
    print("Invalid API key")
except RateLimitError as e:
    print(f"Rate limited. Retry after {e.retry_after}s")
except QuotaExceededError:
    print("Quota exceeded")

Links

• Cloud Dashboard - Sign up & get your API key
• Cloud API Docs - Full API reference
• OSS Repository - Self-hosted option
• Discord - Community & support

License

Apache 2.0